Disclosure: This page may contain affiliate links. Prices were checked on the date shown and may change.
An introduction to VOICEVOX VOICEVOX, based on information from the maker and distributors: what it does, how to use it and who it may suit. It is not scored.
What is VOICEVOX?
VOICEVOX is a free text-to-speech and singing-voice synthesis application. The official site describes it as medium-quality text-to-speech and singing synthesis software that costs nothing to use, the kind of tool people use for narration, video voice-overs and character voices.
Its voices are organized by character, such as Shikoku Metan, Zundamon, Kasukabe Tsumugi and Amehare Hau. The official character list gives each a one-line description of its voice quality, and some characters offer styles beyond the default one.
The official site lists these features: free for commercial and non-commercial use, ready to use right after installation, detailed intonation adjustment, and a humming function that lets the speaking voice sing. Whether commercial use is allowed depends on each character's own terms.
It runs on Windows, Mac and Linux, and the site currently shows version 0.26.1. Besides the editor, the project publishes a speech synthesis engine for use from other applications and a core library for embedding, both open source.
Many character voices
The official list shows around forty characters, each with a short description of the voice. Each has its own terms of use, so check the terms of the voice you choose.
Detailed intonation editing
After entering text you can fine-tune the intonation, which is useful when a reading or accent is off and you want to steer it toward the delivery you want.
Humming / song function
The site states that the speaking voice can sing through a humming function. It is meant for those who want to try melody-based guide-vocal style uses.
Reading and accent dictionary
According to the Q&A, you can register words in the dictionary to correct readings, and export or import it. It helps with scripts full of proper nouns.
CPU and GPU builds
The Q&A says the CPU build supports Windows 10/11, macOS 14+ and Ubuntu 22.04/24.04, while the GPU build supports Windows with a DirectML GPU and Linux with an Nvidia GPU. The Mac build has no GPU mode.
Open engine and core
The synthesis engine is distributed as executables and a Docker image so other software can generate speech. It is an entry point for people building their own tools.
Built on an OSS base
The site explains that VOICEVOX is built on the open-source version; the editor uses Electron and Vue, the engine Python and the core Rust.
Specifications
| Item | Details |
|---|---|
| Price | Free (each character has its own terms) |
| OS (CPU build) | Windows 10/11, macOS 14+, Ubuntu 22.04/24.04 |
| OS (GPU build) | Windows with DirectML GPU, Linux with Nvidia GPU |
| Version | 0.26.1 (shown on the official site) |
| Main functions | Text-to-speech, intonation editing, humming (song) |
| Credit | A credit showing VOICEVOX was used is required (summary of the terms) |
What is included and system requirements
- Editor application (Windows/Mac/Linux)
- Multiple character voices
- Speech synthesis engine and core library (published separately)
- CPU and GPU builds (GPU on Windows/Linux only)
How to use it
- Download the build for your OS from the official site.
- Choose the CPU build if you have no compatible GPU or use a Mac.
- Install and launch the application.
- Pick a character and style.
- Type the text to be read.
- Play it back and adjust accent and intonation as needed.
- Register words in the dictionary to fix misreadings.
- Before using exported audio, check the character's terms and the credit requirement.
Video narration
You could feed in a script and generate a voice for explainer or introduction videos without recording your own voice.
Guide vocals
The humming function might be used to check the feel of a melody during production, as a working reference rather than a final vocal.
Character voices
You might pick a voice that suits a character in streams or videos. Terms differ depending on commercial use, so check them.
Tool integration
The published engine may let you generate speech from your own tools or other applications.
Who it may suit
- People who want to try speech synthesis at no cost
- Those looking for character voices for videos or games
- Users who want to hand-edit readings and accents
- Developers who want to call speech synthesis from their own tools
Good to know before you buy
- Each character has its own terms. Check the terms of the voice you use for commercial-use conditions.
- The summary of the terms says a credit showing that VOICEVOX was used is required.
- The Mac build has no GPU mode, and choosing GPU mode on a PC without a compatible GPU causes errors.
- The Q&A says long text may fail to generate from lack of memory, especially in GPU mode.
- Redistribution and reverse engineering of the software are prohibited by the terms.
Price
Please check the official page for the price.
Free to use (each character has its own terms).
See VOICEVOX on the product page
Sources: voicevox.hiroshiba.jp (checked 2026-10-27)
Comments